Statistically significant. (statistical software packages)
by David Abels

Most people would probably say that statistics has no place in the real world of inventories, sales, and bottom lines. But what if the numbers derived from data you already have on hand could help you get more return on your investment dollars?

"In a lot of small businesses, all the computer does is regurgitate financial data," says Micki Johnson, a business faculty coordinator at Nova University in Fort Lauderdale, Florida. Johnson, who consults with small businesses, encourages proprietors to analyze business data more creatively.

PC-based integrated statistical software is here to help: Minitab ($695; Minitab, 3081 Enterprise Drive, State College, Pennsylvania 16801; 800-448-3555), StatPac Gold IV ($795; StatPac, 3814 Lyndale Avenue South, Minneapolis, Minnesota 55409; 612-822-8252), SPSS/PC+ 4.0 ($195 for first module; SPSS, 444 North Michigan Avenue, Chicago, Illinois 60611; 800-543-2185), SPSS for Windows ($395 for first module), and SYSTAT and SYSTAT for Windows (both $895 for the basic system' SYSTAT, 1800 Sherman Avenue, Evanston, Illinois 60201; 708-864-5760). For the first time, PC users enjoy statistical-analysis power once reserved for corporate mainframe users.

What can you use statistics for? No matter what your type of business, it's important to know who your customers are and what they are thinking. Market research gives you that ability. By surveying a sample of potential or actual customers, you can better determine whether your new product line will be a hot seller.

Market research results are sometimes meaningless, and this might not be obvious from the raw data. You need some measure of the significance of the results. To compute the significance of a survey on a spreadsheet would be cumbersome at best; a statistical package would handle it much more efficiently.

Statistical software also makes it possible to find out how well your customers like what you are currently selling. In fact, you might be able to analyze existing information you have in a customer or accounting database or spreadsheet.

If you are an air-conditioning service company, for example, you may want to know your catchment, or the geographic area in which your customers fall. To figure your catchment, you can draw a scatter diagram based on customer ZIP codes showing your market penetration in various geographical areas. By pulling this information from your accounting database and importing it into a statistical package, you can determine whether, for example, advertising in more local periodicals would serve your needs better than an ad in the city newspaper. Once you've placed your ads, if you've coded them you'll be able to accurately evaluate their effectiveness.

Statistics can help small manufacturers with quality control. Say you publish software. If you want to keep track of how well your staff is packing the materials in the boxes, you can regularly take a sample of your product, enter the independent variables, such as day of the week, hour of the day, number of workers, and rate of production, and the dependent variable: the number of mistakes made. Run a statistical analysis, and you'll be able to determine which of these most powerfully affect shipment quality.

Wouldn't it be interesting to know that when a certain manager or worker is on vacation or sick, productivity and quality improve? Or that certain workers always take sick leave on Monday? A statistical analysis could help you determine where the deadwood is--and where the rotten apples are--in your organization.

While a college course in statistics (dreaded as it might be) can help you to interpret the results obtained from these statistical packages, it's not a requirement. The beauty of these packages is that they enable virtually any small business manager to use statistical methods. Whatever the results, using a statistical software package can lead to strategic planning and a better bottom line for your business.